Anal. Chem. 1998, 70, 3708-3711
Selective Removal of Cesium from Acid Solutions with Immobilized Copper Ferrocyanide T. D. Clarke and C. M. Wai*
Department of Chemistry, University of Idaho Moscow, Idaho 83844
Copper ferrocyanide can be immobilized on Chelex-20, a technical grade chelating resin, for selective removal of cesium from neutral to acidic solutions. The immobilized copper ferrocyanide is much easier to handle than the copper ferrocyanide powder in extraction and separation of cesium from aqueous solutions. In the presence of a reducing agent such as hydrazine, the immobilized copper ferrocyanide resin is able to remove 98% of the cesium from a simulated acid waste solution in one cycle of batch experiments. The resin shows promising properties for remediation of acidic nuclear wastes and for concentration of 137Cs from environmental samples for quantitative analysis. Management of radioactive liquid wastes produced from reprocessing of spent nuclear fuel requires knowledge on methods of radionuclide separation. A significant fraction of the radioactivity in nuclear fuel reprocessing wastes is due to 137Cs, which decays by β and γ emission to stable 137Ba with a half-life of 30.2 years.1 Removing long-lived radionuclides, including 137Cs, from nuclear wastes would result in a significant volume reduction and, consequently, cost saving for their final storage. Methods for 137Cs separation reported in the literature include solvent extraction, precipitation, and ion exchange.1 Solvent extraction of cesium from acid solutions has been demonstrated using selective extractants such as crown ethers or cobalt dicarbollide.2,3 The solvent extraction methods have several drawbacks, including requirement of undesirable solvents, insufficient selectivity in complex matrixes, commercial unavailability of the extractants in large quantities, and unknown radiation stability. Direct decontamination of 137Cs from aqueous solutions using inorganic ionexchangers can avoid some of the problems associated with the solvent extraction processes. Metal ferrocyanides are well-known for their selective and efficient adsorption of cesium.4 The materials are easy to obtain and economical. Structural investigations of metal ferrocyanides have shown that the ferrous ion is surrounded by six octahedrally (1) Schulz, W. W.; Bray, L. A. Sep. Sci. Technol. 1987, 22, 191. (2) Beklemishev, M. K.; Wai, C. M. In Separation Techniques in Nuclear Waste Management; Carleson, T. E., Chipman, N. A., Wai, C. M., Eds.; CRC Press: Boca Raton, FL, 1996; Chapter 3, pp 47-68. (3) Rais, J.; Plesek, J.; Selucky, P.; Kyrs, M.; Kadlecova, L. J. Radioanal. Nucl. Chem. 1991, 148, 349. (4) Ganzerli Valentini, M. T.; Stella, R.; Cola, M. J. Radioanal. Nucl. Chem. 1986, 102, 99.
3708 Analytical Chemistry, Vol. 70, No. 17, September 1, 1998
disposed cyanide groups.5 The octahedra, which are in contact at their corners, form a face-centered cubic lattice. With divalent metal ions, the centers of these cubes are occupied by other cations to balance the net negative charge. The cation may be an alkali metal, a hydrogen ion, or another metal not combined in the complex anion. For example, copper ferrocyanide may possess the limiting compositions CuII[CuIIFe(CN)6], K2[CuIIFe(CN)6], H2[CuIIFe(CN)6], etc., whereas more frequently mixtures are formed such as CuII4K2[CuIIFe(CN)6]5. The cation outside the complex is bound by relatively weaker forces and may be exchanged for another. In most cases investigated, ion-exchange mechanisms have been proposed for cesium adsorption. Most of the metal ferrocyanides studied are excellent sorbents for cesium, with adsorption increasing with the ferrocyanide series: Fe < Cu < Zn < Ni < Co.6 Cesium was readily desorbed by nitric acid from zinc and copper ferrocyanides, while little was desorbed from nickel and cobalt ferrocyanides because of their high stability. Copper and cobalt ferrocyanides are also reported to be resistant to γ radiation.7 Metal ferrocyanides are usually available in very fine powders which are difficult to separate from aqueous solutions by filtration. The metal ferrocyanide powders also have too low a permeability to be of any use for in-column work.8 In view of their undesirable mechanical properties, it appears that immobilizaion of metal ferrocyanides on suitable solid supports would improve such properties and enhance their usefulness in cesium separation. We chose Chelex-20, a chelating ion-exchange resin as the support system for this investigation. One reason is that Chelex-20 is a technical grade 20-50 mesh resin suitable for filtration or column separation. The resin contains iminodiacetic acid groups in a styrene-divinylbenzene copolymer matrix. The selectivity of Chelex-20 for cations in nitrate or chloride media follows the order Cu(II) . Pb(II) > Fe(III) > Al(III) > Cr(III) > Ni(II) > Zn(II) > Co(II) > Cd(II) > Ba(II) > Ca(II) . Na(I).9 Due to the strong bonding of Chelex-20 for Cu(II), we chose to immobilize copper ferrocyanide on this resin for cesium adsorption studies. Experiments were performed in aqueous solutions and in a simulated acid waste solution obtained from the Idaho National Engineering and Environmental Laboratory (INEEL). (5) Huckle, W. Structural Chemistry of Inorganic Compounds; Elsevier: Amsterdam, 1950; Vol. 1. (6) Ganzerli Valentini, M. T.; Meloni, S.; Maxia, V. J. Inorg. Nucl. Chem. 1972, 34, 1427. (7) Pekarek, V.; Vesely, V. Talanta 1972, 19, 1245. (8) Tusa, E.; Paavola, A. Waste Manage. 1993, 1693. (9) Bulletin 1224, Bio-Rad Laboratories, Hercules, CA, 1986. S0003-2700(97)01138-4 CCC: $15.00
© 1998 American Chemical Society Published on Web 07/24/1998
EXPERIMENTAL SECTION Preparation of Immobilized Copper Ferrocyanide. A 35-g quantity of Chelex-20, obtained from Bio-Rad Laboratories (Hercules, CA), was washed with deionized water and 0.1 M HCl. The washed resin was placed in a 250-mL beaker with 100 mL of 1 M Cu(NO3)2 (J.T. Baker, Phillipsburg, NJ) and left to stir overnight. The Cu(II) bonded resin was filtered from the solution and rinsed with deionized water until the washings were clear. The Cu(II) bonded resin was placed in a 100-mL beaker with 10 mL of deionized water. Ten milliliters of 0.25 M K4Fe(CN)6 (Sigma Chemical Co., St. Louis, MO) was added dropwise to the water/ resin slurry with stirring. One or two additional 20-mL aliquots of 0.25 M K4Fe(CN)6 were added dropwise until no blue Cu(II) resin remained. The resulting potassium copper ferrocyanide immobilized resin was filtered from the solution and washed several times with deionized water. The resin was finally washed twice with 10 mL of 0.1 M HCl, each followed by a deionized water rinse. General Physical Characteristics of the Immobilized Resin. The immobilized copper ferrocyanide resin is reddish brown in color. The exact chemical composition of the immobilized material is not known. Our experiments indicate that both potassium and copper are released when the resin is in contact with acidic solutions containing cesium. The ratio of K:Cu released in 1.8 M HNO3 is approximately 5:1, suggesting the possibility of degradation of some copper ferrocyanide in the resin. Thus, a mixed potassium copper ferrocyanide composition is assumed, represented by KCuFCN in this paper. Preliminary investigations showed that the immobilized KCuFCN dissociated in alkaline solution. Nearly all of the immobilized KCuFCN was decomposed from the resin in 4 M LiOH solution. In highconcentration nitric acid solutions (>2 M), the oxidation of ferrocyanide to ferricyanide was observed as evidenced by a visible color change of the resin from reddish brown to a yellowish brown. A similar color change associated with the oxidation of copper ferrocyanide was reported in the literature.6 When the oxidized resin was placed in a 0.05 M hydrazine sulfate (J.T. Baker) solution, the resin was quickly restored to its reddish brown color. Cesium Adsorption Experiments. Radioactive 137Cs was used as a tracer for this study. The radioisotope was obtained from Isotope Product Laboratories (Burbank, CA). The decay scheme of 137Cs is shown as follows:
The γ peak at 662 keV emitted from the decay of 137mBa was used to quantify 137Cs. The characteristic γ ray was measured using a Ge(Li) detector (EGG-ORTEC) combined with a 4096 multichannel analyzer (ORTEC). A trace amount of 137Cs was added to 10 mL of a 1 mM unlabeled cesium solution prepared with CsCl (Sigma Chemical Co.), diluted with deionized water, 1.8 M HCl, or 1.8 M HNO3. A
Figure 1. Logarithm of the distribution coefficient (Kd) plotted against the negative logarithm of initial Cs concentration ([Cs]o) for cesium adsorption on the KCuFCN immobilized resin. (9) Water, ([) 1.8 M HCl, and (b) 1.8 M HNO3.
1-mL sample was taken, and the initial activity (Ao) of the aliquot was measured. After counting, this sample was returned to the original solution, and 0.1 g of the immobilized KCuFCN was added to the solution. After the solution was shaken for a period of time, another 1-mL aliquot was removed from the solution. The solution was left undisturbed for about 25 min for 137mBa to reach secular equilibrium with the parent, and then the activity (At) was measured again using the γ spectrometer. Generally speaking, after 30 min of shaking, the activity of 137Cs in the solution no longer showed an appreciable change. To ensure equilibrium, the sample was shaken for 1 h. The distribution coefficients (KD) were calculated as KD ) activity of 137Cs sorbed per gram of adsorbent/activity of residual 137Cs in solution per milliliter. The residual concentration of cesium in the solution (Ce) was calculated from the equation Ce ) (At/Ao)[Cs]o, where [Cs]o is the original concentration of cesium. The amount of cesium adsorbed per unit weight of the solid adsorbent (qe) was calculated from the equation qe ) (Ao - At/Ao)(mg of [Cs]o/g of KCuFCN). RESULTS AND DISCUSSION The immobilized KCuFCN is able to adsorb Cs+ in aqueous solutions effectively from neutral pH to about 2 M acid solutions. Figure 1 shows the log KD values plotted versus the initial cesium concentration (also in logarithm scale) for aqueous solutions at pH 6.5 and with 1.8 M HCl or 1.8 M HNO3. At initial cesium concentrations of less than 3 × 10-3 M, the log D for Cs reaches a value slightly above 4 for water at pH 6.5 and for the 1.8 M HCl solution. In 1.8 M HNO3, the log D value is about 2.6 at [Cs]o ) 3 × 10-3 M. The lower D value observed in the nitric acid solution is probably related to the oxidation of iron in KCuFCN. The percent cesium adsorption by KCuFCN as a function of time was measured in both 1.8 M HCl and 1.8 M HNO3 (Figure 2). These curves show the fast initial adsorption characteristic of inorganic exchangers with 98% removal of cesium from both acidic solutions after 1 h of shaking. When the KCuFCN was left in the presence of HNO3, a decrease in the percent of cesium sorbed was observed, resulting in only 90% adsorption. This desorption process is attributed to the oxidation of some of the copper ferrocyanide to copper ferricyanide by HNO3 resulting in the release of cesium. Similar oxidation behavior was observed Analytical Chemistry, Vol. 70, No. 17, September 1, 1998
3709
Table 1. Composition of the Simulated Liquid Waste element
concn (M)
element
concn (M)
Na Ca Mn Pb Cd Ni Hg Al B Cl PO43-
1.25 4.40 × 10-2 1.40 × 10-2 1.00 × 10-3 2.00 × 10-3 2.00 × 10-3 2.00 × 10-3 5.48 × 10-1 1.60 × 10-2 2.20 × 10-2 1.00 × 10-2
K Fe Mo Cr Zr Cs Sr Ce F SO42NO3-
1.44 × 10-1 2.50 × 10-2 1.00 × 10-3 6.00 × 10-3 5.00 × 10-3 6.77 × 10-5 1.80 × 10-5 4.00 × 10-4 7.10 × 10-2 3.80 × 10-2 4.49
Figure 2. Relative adsorption rate of Cs on the KCuFCN immobilized resin from (2) 13 ppm Cs in 1.8 M HCl, (9), 13 ppm Cs in 1.8 M HNO3, and (b) 133 ppm Cs in 1.8 M HNO3. a
In 1.8 M HNO3.
Table 2. Percent Cesium Adsorption by KCuFCN Resin from Simulated Liquid Waste with Added Reducing Agentsa
Figure 3. Effect of hydrazine on the adsorption of Cs in 1.8 M HNO3 on the KCuFCN immobilized resin. (O) 133 ppm Cs in 5 mM hydrazine, (b) 133 ppm Cs without hydrazine, (0) 13 ppm Cs in 5 mM hydrazine, and (9) 13 ppm Cs without hydrazine.
for zinc ferrocyanide.3 The following redox process was proposed by Ganzerli Valentini et al.:4
3FeII(CN)64- + 4H+ + NO3- f 3FeIII(CN)63- + NO + 2H2O This oxidation process can be prevented using a reducing agent such as hydrazine. Figure 3 shows the adsorption of cesium by KCuFCN with the addition of hydrazine sulfate. The oxidation of the ferrocyanide in 1.8 M HNO3 appears to be inhibited as the desorption of cesium is no longer observed and percent adsorption curves in the HNO3 solution are comparable to the HCl solution. The capacity of the KCuFCN resin for Cs in 1.8 M HNO3 was evaluated by plotting 1/qe versus 1/Ce of the experimental data. A linear relationship was observed, indicating that Cs adsorption follows a simple Langmuir model. The intercept of 1/qe at 1/Ce ) 0 gave a qo (capacity) value of about 1.3 × 10-3 mol of Cs/g of the resin. A simulated acid waste solution was obtained from INEEL with its chemical composition listed in Table 1. The simulated waste solution was in 1.8 M HNO3 and contained high concentrations of sodium, potassium, and other metals. Batch experiments were performed by adding 0.1 g of the KCuFCN resin in 25 mL of the simulated waste solution spiked with a trace amount of 137Cs. The mixture was shaken for 15 h using a wrist-action mechanical shaker (Burrell model 75). The 137Cs activities in the solution before and after shaking were measured using the γ spectrometer described in the Experimental Section. Table 2 shows the results of cesium adsorption from the simulated waste solution under various conditions. As shown in Table 2, only about 60% of the cesium in the simulated acid solution can be removed by the 3710 Analytical Chemistry, Vol. 70, No. 17, September 1, 1998
a
reducing agent
% Cs adsorption
none 2 mM hydrazine 5 mM hydrazine 5 mM ascorbic acid 10 mM ascorbic acid
58 97 98 63 80
0.1 g of KCuFCN resin in 25 mL of the simulated liquid waste.
KCuFCN resin under the specified experimental conditions. This is significantly lower than the results observed in 1.8 M HNO3 solution shown in Figure 1. The presence of various cations plus a high concentration of NO3- in the simulated solution apparently can interfere with the adsorption of cesium by the resin. However, when a reducing agent such as hydrazine was added to the solution, the percent of cesium removal can be significantly enhanced. At 5 × 10-3 M hydrazine, about 98% of the cesium in the acid solution was removed by the resin. Ascorbic acid can also enhance the efficiency of cesium adsorption, but not as effectively as hydrazine. With 2 × 10-2 M ascorbic acid in the acid solution, the adsorption of cesium was increased to 80%. The enhanced adsorption of cesium is probably caused by the reducing agents’ ability to inhibit the oxidation of KCuFCN. The rate of cesium adsorption from the simulated acid solution with 5 × 10-3 M hydrazine is similar to those shown in Figure 3. Over 90% adsorption of cesium was actually achieved after only 30 min of shaking. Similar experiments were also performed to evaluate the removal of other metals from the simulated solution by the KCuFCN resin. In this case, radioactive 137Cs was not used, and the solution was analyzed by ICPMS. The amounts of Cr, Cd, Fe, Hg, Pb, and Zr removed by the KCuFCN resin (0.1 g) with 5 × 10-3 M hydrazine in the simulated acid solution (25 mL) were measured. No detectable amounts of Cr and Cd were removed by the resin, and only
